Duodenal Ulcer-related Antigens from Helicobacter pylori - CiteSeerX

Research

Duodenal Ulcer-related Antigens from Helicobacter pylori S IMMUNOPROTEOME AND PROTEIN MICROARRAY APPROACHES*□

Yu-Fen Lin‡, Chun-Yi Chen‡, Mong-Hsun Tsai§, Ming-Shiang Wu¶储, Yu-Chun Wang**, Eric Y. Chuang‡‡, Jaw-Town Lin¶储, Pan-Chyr Yang¶, and Lu-Ping Chow‡§§¶¶ Helicobacter pylori is an important risk factor of duodenal ulcer (DU). Although many virulence factors of H. pylori have been identified, few have been reported to show an association with the pathogenesis of DU. The aims of this study were to identify H. pylori antigens showing a high seropositivity in DU and to develop a platform for rapid and easy diagnosis for DU. Because DU and gastric cancer (GC) are considered clinical divergent gastroduodenal diseases, we compared two-dimensional immunoblots of an acid-glycine extract of an H. pylori strain from a patient with DU probed with serum samples from 10 patients with DU and 10 with GC to identify DU-related antigens. Of the 11 proteins that were strongly recognized by serum IgG from DU patients, translation elongation factor EF-G (FusA), catalase (KatA), and urease ␣ subunit (UreA) were identified as DU-related antigens, showing a higher seropositivity in DU samples (n ⴝ 124) than in GC samples (n ⴝ 95) (FusA, 70.2 versus 45.3%; KatA, 50.8 versus 41.1%; UreA, 44.4 versus 27.4%). In addition, we found that the use of multiple antigens improved the discrimination between patients with DU and those with GC as the odds ratios increased from 1.82 (95% confidence interval (CI), 0.79 – 4.21; p ⴝ 0.1607) for seropositivity for FusA, KatA, or UreA alone to 4.95 (95% CI, 2.05–12.0; p ⴝ 0.0004) for two of the three antigens and to 5.71 (95% CI, 1.86 –17.6; p ⴝ 0.0024) for all three antigens. Moreover a protein array containing the three DU-related antigens was developed to test the idea of using multiple biomarkers in diagnosis. We conclude that FusA, KatA, and UreA are DU-related antigens of H. pylori, and the combination of these on a protein array provided a rapid and convenient method for detecting serum antibody patterns of DU patients. Molecular & Cellular Proteomics 6:1018 –1026, 2007.

From the ‡Graduate Institute of Biochemistry and Molecular Biology and 储Department of Primary Care Medicine, College of Medicine, National Taiwan University, Taipei 10051, Taiwan, §Institute of Biotechnology, College of Bio-Resources and Agriculture and ‡‡Department of Electrical Engineering, College of Electrical Engineering and Computer Science, National Taiwan University, Taipei 10617, Taiwan, Departments of ¶Internal Medicine and §§Medical Genetics, National Taiwan University Hospital, Taipei 10048, Taiwan, and **Institute of Environmental Health, College of Public Health, National Taiwan University, Taipei 10055, Taiwan Received, January 9, 2007, and in revised form, February 15, 2007 Published, MCP Papers in Press, February 21, 2007, DOI 10.1074/ mcp.M700009-MCP200

1018

Molecular & Cellular Proteomics 6.6

Helicobacter pylori is the major factor involved in the pathogenesis of gastritis, gastric ulcer, duodenal ulcer (DU),1 and gastric cancer (GC). It is known that this bacterium infects more than half the world’s population (1). However, the clinical outcomes of H. pylori infection are highly variable and influenced by both bacterial factors and host immune responses. Most infected persons are asymptomatic with no apparent disease, but 6 –20% have duodenal ulceration, and a small proportion develop GC (2). DU and GC are considered clinically divergent gastroduodenal diseases in which the pattern of gastritis is regarded as the major factor in the pathogenesis. Antral-predominant gastritis leads to increased acid production and duodenal ulceration, whereas corpus-predominant atrophic gastritis leads to acid reduction and a higher risk of developing GC (3, 4). It has been recognized that the inflammation induced by H. pylori infection is a factor related to the pathogenesis of gastroduodenal diseases. Chronic active gastritis, the hallmark of H. pylori infection, is characterized by infiltration of the mucosa by neutrophils, lymphocytes, and monocytes/macrophages (5). A strong humoral immune response to a variety of H. pylori antigens is also elicited (6). H. pylori antigens are therefore regarded as potential candidates for biomarkers or vaccines (7). Serological tests are noninvasive methods for diagnosing H. pylori infection. Moreover evaluation of the humoral immune responses to H. pylori antigens by immunoblotting appears to be more sensitive than ELISA for detecting low abundance antibodies. Importantly ⬃70% of patients with DU are infected with H. pylori (8, 9). The aims of the present prospective study were to identify DU-related antigens by two-dimensional (2D) immunoblotting and mass spectrometry and to develop a platform for detecting antibody patterns for diagnostic use. Protein array technology has been shown to be a useful tool for multiplexed measurements and proteomics studies. Protein arrays can be classified into two types, analytical and functional protein arrays. The most representative analytical protein arrays are antibody arrays. Functional protein arrays have been applied to many fields, including studies of proteinprotein, protein-DNA, and protein-drug interactions; biomarker discovery; and clinical diagnosis (10). The binding of humoral antibodies to allergens, autoantigens, cancer-specific antigens, or infectious organisms is the basis of the design of protein arrays for biomarker discovery and clinical diagnosis 1

The abbreviations used are: DU, duodenal ulcer; 2D, two-dimensional; CI, confidence interval; GC, gastric cancer; OR, odds ratio.

© 2007 by The American Society for Biochemistry and Molecular Biology, Inc. This paper is available on line at http://www.mcponline.org

Duodenal Ulcer-related H. pylori Antigens

(11–14). Applying this concept, we wished to develop a DUrelated protein array, which can be used for immunodiagnostics in the form of miniaturized and multiplexed assays. In this study, we used a proteomics approach to identify DU-related antigens of H. pylori by comparing the profiles of 2D immunoblots of an acid-glycine extract of H. pylori probed with DU and GC sera. We identified three DU-related antigens, FusA, KatA, and UreA, all of which were more frequently recognized by DU sera than GC sera. Statistical analysis of the data showed that the discrimination power of the threeantigen combination was greater than that of a single antigen. We therefore developed a DU-related protein array that can be used to simultaneously and rapidly detect the reactions of serum antibodies to different DU-related antigens. EXPERIMENTAL PROCEDURES

Isolation of the H. pylori Strain and Culture Conditions—H. pylori strain HD12 was isolated from endoscopic biopsy samples from the stomach of a patient with DU at the National Taiwan University Hospital. The bacteria were cultured on a BBLTM StackerTM plate (BD Biosciences) at 37 °C under microaerobic conditions. Patients and Serum Samples—Serum samples were prospectively collected from individuals who participated in a national project for the investigation of H. pylori and gastroduodenal disorders in Taiwan between December 1999 and December 2001. Our study protocol was approved by both the Institutional Research Board and the Department of Health, Executive Yuan, Taiwan. One hundred and twenty-four patients with DU who received an upper gastrointestinal endoscopic examination were enrolled. Ninety-five patients with GC who underwent curative gastrectomy at our institution were also enrolled. H. pylori status was determined by culture and/or histological examination of gastric biopsy specimens. In addition, 40 subjects with a normal appearance of the gastric mucosa and no evidence of H. pylori infection were selected as controls. Fasting serum samples were collected from all participants, catalogued, aliquoted, and stored at ⫺80 °C. Aliquots were only thawed once prior to analysis. Two-dimensional Electrophoresis and Immunoblotting—Cell surface proteins were extracted from H. pylori using an acid-glycine extraction procedure as described previously (15). Proteins in the H. pylori acid-glycine extract were precipitated using trichloroacetic acid (20%) and separated by 2D electrophoresis as described previously (16). Briefly the protein extract was incubated with 2D sample buffer (8 M urea, 2% Pharmalyte, pH 3–10, 60 mM DTT, 4% CHAPS, bromphenol blue), the first dimension of the 2D gel was run on IPG strips (Immobiline DryStrip, pH 3–10, 11 cm, GE Healthcare), and the second dimension was run on 12.5% SDS-polyacrylamide gels. For immunodetection, the proteins on the gel were transferred to a PVDF membrane (Millipore, Bedford, MA). Then the membrane was blocked by incubation for 1 h at room temperature in blocking buffer (26 mM Tris-HCl, 150 mM NaCl, pH 7.5, 1% skimmed milk) and incubated overnight at 4 °C with serum samples from DU patients or GC patients or pooled normal sera (1:1000 in blocking buffer containing 0.05% Tween 20) and then for 1 h at room temperature with horseradish peroxidase-conjugated goat anti-human IgG (Chemicon, Temecula, CA). Then bound antibody was detected using 3-amino-9ethylcarbazole (Sigma) as substrate. Protein Identification—The individual Coomassie Blue-stained protein spots were excised and subjected separately to in-gel tryptic digestion. Briefly the spots were destained using 50 mM NH4HCO3 in 50% acetonitrile and dried in a SpeedVac concentrator. The protein was then digested by incubation overnight at 37 °C with sequencing grade trypsin (Promega, Madison, WI) in 50 mM NH4HCO3, pH 7.8. The resulting peptides were extracted sequentially with 1% TFA and

0.1% TFA, 60% acetonitrile, and the combined extracts were lyophilized and analyzed using a QSTARTM XL Q-TOF mass spectrometer (Applied Biosystems, Framingham, MA) coupled to an UltiMateTM nano-LC system (Dionex/LC Packings, Amsterdam, Netherlands). Peak lists of MS/MS spectra were created using Mascot Search version 1.6b4 in Analyst姞 QS 1.1 (Applied Biosystems) and uploaded to the Mascot MS/MS Ions Search program (Mascot version 2.0) on the Matrix Science public website, and protein identification was performed against the National Center for Biotechnology Information non-redundant (NCBInr) database (containing 3,957,439 protein entries at the time searched). Up to one missed cleavage was allowed. Cysteine carbamidomethylation, glutamine/asparagine deamidation, and methionine oxidation were set as possible modifications. The error windows for peptide and MS/MS fragment ion mass values were 0.3 and 0.5 Da, respectively. MH22⫹ and MH33⫹ were selected as the precursor peptide charge states in the search. Ion scores greater than 54 indicate a significant match; the individual score for the MS/MS spectrum of each peptide was more than 20. From the hit lists, the protein names and locus_tag for H. pylori strain 26695 were selected and are listed in Table I and Supplemental Table 1. Cloning and Purification of the Recombinant Proteins—H. pylori was lysed, the lysate was subjected to RNase treatment, and then genomic DNA was extracted using phenol-chloroform and precipitated with 70% ethanol. To amplify the DNA fragments containing the H. pylori fusA gene, katA gene, ureA gene, and flaA gene (control) by PCR, the primer pairs used were: fusA: sense, 5⬘-GGT ACC ATG GCT AGA AAA ACC CCA-3⬘; antisense, 5⬘-CTG CAG TCA GCC TTT GCG TTT TTC-3⬘; katA: sense, 5⬘-GAA TTC GAT GGT TAA TAA AGA TGT GAA-3⬘; antisense, 5⬘-CTC GAG CTT TTT CTT TTT TGT GTG GT-3⬘; ureA: sense, 5⬘-GGA TCC ATG AAA CTC ACC CCA-3⬘; antisense, 5⬘-GGA TCC TTA CTC CTT AAT TGT T-3⬘; and flaA: sense, 5⬘-GGA TCC ATG GCT TTT CAG GTC AAT-3⬘; antisense, 5⬘-GGT ACC CTA AGT TAA AAG CCT TAA G-3⬘. PCR was performed using 30 cycles of 94 °C for 1 min, 55 °C for 1 min, and 72 °C for 2 min followed by a final extension at 72 °C for 15 min. The fusA, ureA, and flaA gene fragments were cloned into the expression vector pQE30 (Qiagen, Hilden, Germany) and transformed into Escherichia coli strain M15, whereas the katA gene fragment was cloned into the pET-22b(⫹) vector (Novagen, Madison, WI) and transformed into E. coli strain BL21. For expression of recombinant proteins, cells were grown to an A600 value of 0.6, induced with 1 mM isopropyl ␤-D-thiogalactoside, and harvested after 6 h at 25 °C (for UreA), 2 h at 37 °C (for FlaA), or 4 h at 37 °C (for FusA and KatA). The recombinant proteins were dissolved in binding buffer (20 mM Tris-HCl, 0.5 M NaCl, 5 mM imidazole, pH 7.9) containing 8 M urea and purified on a Ni2⫹-chelating Sepharose column (GE Healthcare). Serologic Study—An H. pylori acid-glycine extract and recombinant FusA, KatA, UreA, and FlaA were electrophoresed on 12.5% SDS-polyacrylamide gels, transferred to the PVDF membrane, and then subjected to immunoblotting using serum samples from patients with DU or GC or from normal controls diluted 1:1000 as described above. Statistical Analysis—Statistical analysis was performed using SAS version 9.1 (SAS Institute Inc., Cary, NC). The odds ratio (OR) and 95% confidence interval (CI) were estimated by multiple logistic regression analysis. Comparisons between tests of serum reactions on protein arrays were made using Student’s t test. A p value of ⬍0.05 was considered statistically significant. Protein Array Fabrication—Poly(L-lysine)-coated slides were placed in an OmniGrid arrayer (GeneMachines, San Carlos, CA) and four SMP3 printing pins (TeleChem, Sunnyvale, CA) with a tip diameter of 100 ␮m were used to generate the protein arrays. All proteins were spotted in quadruplicate at three different concentrations (100, 300, and 600 ␮g/ml) diluted with 0.14 –2 M urea in PBS. Each protein array


1019


contained the control spots, human IgG in three dilutions (one normal serum was diluted with PBS to 1:1500, 1:1000, and 1:500). The resulting protein chips were stored in a desiccator. Serum Reaction on Protein Arrays—All reactions were performed at room temperature. The chips were blocked for 2 h with 2% BSA in TBST (20 mM Tris, 137 mM NaCl, 0.1% Tween 20, pH 7.6) in a humidified atmosphere and then incubated with the test sera (1:100 in 2% BSA in TBST) for 1 h. Following three TBST washes, the chips were incubated for 1 h in the dark with Cy3-conjugated anti-human IgG antibody (1:500 in 2% BSA in TBST; Abcam, Cambridge, UK) and then washed three times in TBST and twice in distilled water. All incubation steps were performed in a volume of 50 ␮l underneath a cover slide. Following air drying, the protein chips were imaged using a GenePixTM 4000B scanner (Axon Instruments, Union City, CA), and

FIG. 1. One-dimensional immunoblotting of an H. pylori extract using serum samples. An acid-glycine extract of cell surface proteins from H. pylori was separated by 12.5% SDS-PAGE, and then the separated proteins were transferred to a PVDF membrane that was stained with Coomassie Blue (A, lane 1) or immunoblotted with sera from H. pylori-infected patients with DU (A, lanes 2–11) or H. pylorinegative normal individuals (B, lanes 1–10).

image analysis was performed using GenePix Pro 6.0 (Axon Instruments). RESULTS

Identification of DU-related Antigens of H. pylori by 2D Immunoblot Analysis—To identify the DU-related antigens of H. pylori, we first examined the one-dimensional immunoblot profiles of the acid-glycine-extracted bacterial proteins probed with sera from H. pylori-infected patients with DU. The molecular masses of the immunoreactive proteins ranged from 23 to 97 kDa (Fig. 1A). The binding pattern of each serum sample was distinct and unique. Moreover several proteins were recognized by sera from normal individuals (Fig. 1B). We then performed 2D electrophoresis on the bacterial proteins and examined the patterns of the 2D immunoblots probed with sera from 10 patients with DU in the active stage. Fig. 2A shows the complex 2D profile of the H. pylori acid-glycineextracted proteins after silver staining. The representative 2D immunoblot probed with one DU serum is shown in Fig. 2B. Numerous protein spots recognized by the DU sera were observed. The majority of the recognized antigens had molecular masses greater than 30 kDa under reducing conditions, and those showing the strongest reaction had molecular masses greater than 50 kDa. To determine which antigens were DU-related and non-DU-related, we compared the frequency of recognition of these by sera from patients with DU and GC (n ⫽ 10) and identified the proteins by nano-LCMS/MS analysis (Table I and Supplemental Table 1). The representative 2D immunoblot probed with one GC serum is shown in Fig. 2C. Several proteins recognized at a high frequency by both DU and GC sera were flagellar hook protein, flagellar hook-associated protein (FliD), molecular chaperone

FIG. 2. 2D profiles of immunoreactive proteins in the H. pylori sample. The acid-glycine extract of cell surface proteins from H. pylori was separated by 2D electrophoresis using a linear pH 3–10 gradient in the first dimension and 12.5% SDS-PAGE in the second dimension. The separated proteins were detected by silver staining (A) or were transferred to a PVDF membrane and probed with serum from an H. pylori-infected patient with DU (B) or GC (C).

1020



TABLE I Frequency of seropositivity for the indicated H. pylori antigens in the DU and GC groups identified by nano-LC-MS/MS analysis Seropositivity Theoretical Sequence Locus_tag pI/molecular Score DU GC coverage mass (Da) (n ⫽ 10) (n ⫽ 10)

Spot no.

Protein

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Translation elongation factor EF-G (FusA) Flagellar hook protein Flagellar hook-associated protein (FliD) Chaperone protein DnaK (Actinobacillus actinomycetemcomitans) Molecular chaperone DnaK Urease ␤ subunit (UreB) Chaperonin GroEL Flagellin B (FlaB) Flagellin A (FlaA) Catalase (KatA) Flagellin A (FlaA) Serine protease (HtrA) Urease ␣ subunit (UreA) Urease ␣ subunit (UreA) Urease ␣ subunit (UreA)

%

DnaK, urease ␤ subunit, flagellin A, flagellin B, and serine protease (HtrA). Importantly comparison of DU and GC prevalence data showed that translation elongation factor EF-G (FusA; DU versus GC, 70 versus 40%), catalase (KatA; 100 versus 50%) and urease ␣ subunit (UreA; 40 and 80% versus 20%) were potential candidates for DU-related antigens. Fig. 3 shows the reactivity of FusA, KatA, and UreA in 2D immunoblots probed with 10 DU sera. The location of FusA was close to a spot with strong immunoreactivity. The different isoforms of KatA were recognized equally well by DU sera; in contrast, 40% of DU sera recognized two of three UreA isoforms, and 40% of DU sera recognized all UreA isoforms. Clinical Significance of FusA, KatA, and UreA Seropositivity—To further examine the clinical significance of FusA, KatA, and UreA as potential candidates for DU-related antigens, we expressed the recombinant His-tagged fusion proteins in E. coli. Recombinant FusA, KatA, and UreA with apparent molecular masses of 80, 63, and 28 kDa, respectively, were successfully expressed and purified (Fig. 4, lanes 1–3). The identity of the purified proteins was confirmed by nano-LCMS/MS, and their antigenicity was examined by immunoblot analysis using pooled sera from DU patients (Fig. 4, lanes 4 – 6). They were then used in an immunoblot assay using a series of clinical samples including sera from H. pylori-infected patients with DU (n ⫽ 124) or GC (n ⫽ 95) or non-H. pylori-infected normal individuals (control, n ⫽ 40). As shown in Table II, the prevalence of FusA seropositivity was significantly higher in DU patients than in GC patients (70.2 versus 45.3%; OR ⫽ 2.68; 95% CI, 1.52– 4.72; p ⫽ 0.0007) or controls (70.2 versus 20%; OR ⫽ 8.42; 95% CI, 3.23–22.0; p ⬍ 0.0001). The corresponding values for KatA seropositivity in patients with DU, GC, and control were 50.8, 41.1, and 5%. KatA seropositivity in DU patients was significantly higher than in controls (OR ⫽ 17.7; 95% CI, 3.83– 81.8; p ⫽ 0.0002) but not significantly different from that in GC patients (OR ⫽

HP1195 HP0870 HP0752 HP0109 HP0072 HP0010 HP0115 HP0601 HP0875 HP0601 HP1019 HP0073 HP0073 HP0073

5.25/76,972 5.04/76,160 5.15/74,043 4.88/68,352 5.04/67,011 5.71/61,676 5.44/58,236 5.95/53,923 5.58/53,235 8.57/58,599 5.58/53,235 9.18/51,593 8.47/26,537 8.47/26,537 8.47/26,537

5 36 34 28 36 52 52 29 43 40 5 56 46 66 61

%

68 292 183 436 330 788 500 153 575 259 72 443 144 288 313

70 100 100 90 100 90 70 100 100 100 100 100 40 80 80

40 100 80 80 70 80 80 100 100 50 100 80 20 20 20

1.45; 95% CI, 0.82–2.56; p ⫽ 0.1976). UreA seropositivity in DU patients was significantly higher than in GC patients (44.4 versus 27.4%; OR ⫽ 2.05; 95% CI, 1.13–3.71; p ⫽ 0.0183) or controls (44.4 versus 15%; OR ⫽ 4.43; 95% CI, 1.51–13.1; p ⫽ 0.0069). Moreover the seropositivity of each protein in DU patients was significantly different from that in the GC patients and normal individuals combined (FusA: OR ⫽ 3.37; 95% CI, 1.97–5.76; p ⬍ 0.0001; KatA: OR ⫽ 2.16; 95% CI, 1.25–3.73; p ⫽ 0.0055; UreA: OR ⫽ 2.38; 95% CI, 1.35– 4.20; p ⫽ 0.0029). As a result, H. pylori FusA, KatA, and UreA were antigens related to DU disease. Combination of Multiple Antigens as Biomarkers of DU Disease—To evaluate whether a combination of multiple antigens increased the ability to distinguish patients with DU from those with GC and normal individuals, we compared the frequency of seroreactivity with a single antigen (FusA, KatA, or UreA) and with combinations of these antigens in gastroduodenal diseases (Table III). Comparing the results for the DU samples with those for the GC samples, the odds ratios increased with the number of recognized antigens, going from 1.82 (95% CI, 0.79 – 4.21; p ⫽ 0.1607) for a single recognized antigen to 4.95 (95% CI, 2.05–12.0; p ⫽ 0.0004) for two recognized antigens and to 5.71 (95% CI, 1.86 –17.6; p ⫽ 0.0024) for three recognized antigens. Similarly when the results for the DU samples were compared with those for the GC samples and normal samples combined, the odds ratios again increased with the number of recognized antigens: 3.14 (95% CI, 1.46 – 6.77; p ⫽ 0.0035) for a single antigen, 9.87 (95% CI, 4.35–22.4; p ⬍ 0.0001) for two antigens, and 12.5 (95% CI, 4.23–36.8; p ⬍ 0.0001) for three antigens. Using combinations of two of the three antigens, each pair had a significantly higher odds ratio than the single antigen when DU samples were compared with GC samples alone or the population of GC and normal individuals combined. Importantly the combination of KatA and UreA led to the best


1021


FIG. 4. Immunoreactivity of the recombinant proteins. Purified recombinant FusA (lanes 1 and 4), KatA (lanes 2 and 5), and UreA (lanes 3 and 6) were electrophoresed on a 10% SDS gel and detected by Coomassie Blue (lanes 1–3) or transferred to a PVDF membrane and immunoblotted with pooled DU sera (lanes 4 – 6). Lane M, molecular mass markers.

FIG. 3. The patterns of DU-related antigens on 2D immunoblots. The acid-glycine extract of cell surface proteins from H. pylori was separated by 2D electrophoresis and transferred to the PVDF membrane; the portions of the silver-stained gel and the immunoblots containing FusA, KatA and UreA protein spots are shown (indicated by arrows). The 2D immunoblots were analyzed by blotting with sera from 10 DU patients. WB, Western blot.

discrimination of DU disease from GC (OR ⫽ 16.0; 95% CI, 1.77–144; p ⫽ 0.0136) or a population of GC and normal individuals (OR ⫽ 34.9; 95% CI, 3.95–308; p ⫽ 0.0014). According to the statistical data, we can use two-antigen and three-antigen combinations of FusA, KatA, and UreA as biomarkers of DU. Development of a DU-related Antigen Protein Array—To improve the efficiency and convenience of diagnostic usage, we designed a multiple protein array including DU-related

1022


antigens and a non-disease-related antigen in a poly(L-lysine)coated slide. The DU-related antigens were FusA, KatA, and UreA, whereas FlaA was selected as a positive control antigen to monitor the experimental performance. We found that FlaA was recognized by sera from H. pylori-infected persons and non-infected persons. All antigens were spotted at three different concentrations (100, 300, and 600 ␮g/ml) to show the trend of antigen-antibody recognition. Human IgG was also included at three serial dilutions for interchip comparison. The chips were incubated with sera from six DU patients and seven normal individuals. Based on the immunoblot analysis, six DU patients whose serum IgG recognized all the DUrelated antigens were selected. In contrast, none of the DUrelated antigens were recognized by serum IgG from the seven normal individuals. The results of chip reactions with six DU sera and four normal sera are shown in Fig. 5A. The analysis of human IgG showed a low coefficient of variation (26.4, 28.9, and 3.6% for 1:1500, 1:1000, and 1:500, respectively), indicating good reproducibility between chip-serum incubations. When comparing DU sera with normal sera using the average intensity of fluorescence, obviously distinct two groups were observed at each concentration of each protein (Fig. 5B). This experiment demonstrated that the diagnosis testing of a combination of antigens could be performed using a protein array. DISCUSSION

H. pylori is an important factor in the pathogenesis of DU (17). Clinically about 70% of endoscopically diagnosed DU patients are infected with H. pylori (8, 9). To identify potential candidate H. pylori antigens related to DU, we used a proteomics approach to analyze H. pylori antigens from a strain clinically isolated from a patient with DU in the active stage and found 11 proteins, which reacted strongly with sera from DU patients. Most of these were recognized by all of the DU


TABLE II Frequency of seroreactivity with a single recombinant antigen in patients with gastroduodenal diseases or controls and the odds ratio OR (95% CI), p valuea

Positive no. (%)

a

Antigens

DU (n ⫽ 124)

GC (n ⫽ 95)

Control (n ⫽ 40)

DU vs. GC

DU vs. control

DU vs. (GC ⫹ control)

FusA KatA UreA

87 (70.2) 63 (50.8) 55 (44.4)

43 (45.3) 39 (41.1) 26 (27.4)

8 (20) 2 (5) 6 (15)

2.68 (1.52–4.72), 0.0007 1.45 (0.82–2.56), 0.1976 2.05 (1.13–3.71), 0.0183

8.42 (3.23–22.0), ⬍0.0001 17.7 (3.83–81.8), 0.0002 4.43 (1.51–13.1), 0.0069

3.37 (1.97–5.76), ⬍0.0001 2.16 (1.25–3.73), 0.0055 2.38 (1.35–4.20), 0.0029

The OR, 95% CI, and p value were obtained by multiple logistic regression analysis. TABLE III Effect of an increased number of recognized antigens on the odds ratio Recognized antigens No.

Antigens

DU (n ⫽ 124)

GC (n ⫽ 95)

GC ⫹ control (n ⫽ 135)

DU vs. GC

DU vs. (GC ⫹ control)

0 1 2

None Any one antigen Two of the three antigens FusA and KatA KatA and UreA FusA and UreA All three antigens

11 (8.9) 41 (33.1) 52 (41.9) 26 (21.0) 8 (6.5) 18 (14.5) 20 (16.1)

22 (23.2) 45 (47.4) 21 (22.1) 12 (12.6) 1 (1.1) 8 (8.4) 7 (7.4)

48 (35.6) 57 (42.2) 23 (17.0) 12 (8.9) 1 (0.7) 10 (7.4) 7 (5.2)

1.0 1.82 (0.79–4.21), 0.1607 4.95 (2.05–12.0), 0.0004 4.33 (1.60–11.7), 0.0039 16.0 (1.77–144), 0.0136 4.50 (1.49–13.6), 0.0075 5.71 (1.86–17.6), 0.0024

1.0 3.14 (1.46–6.77), 0.0035 9.87 (4.35–22.4), ⬍0.0001 9.45 (3.67–24.4), ⬍0.0001 34.9 (3.95–308), 0.0014 7.85 (2.85–21.6), ⬍0.0001 12.5 (4.23–36.8), ⬍0.0001

3 a

OR (95% CI), p valuea

Positive no. (%)

The OR, 95% CI, and p value were obtained by multiple logistic regression analysis.

sera used in the 2D immunoblot analysis in our study and most have been reported as immunoreactive proteins in previous studies (18 –20). Kimmel et al. (18) identified 29 antigens that reacted with sera from patients with different gastroduodenal diseases but failed to show a correlation between antibody patterns in patients and particular diseases. Haas et al. (19) probed H. pylori proteins from the strain 26695 with sera from patients with active H. pylori infection suffering from gastritis, ulcer, or gastric cancer. They identified 32 H. pylori antigens reactive with H. pylori-positive sera and characterized several antigens differentially recognized by sera from gastritis and ulcer patients. Mini et al. (20) analyzed the H. pylori proteins of three strains isolated from patients with chronic gastritis, duodenal ulcer, and GC and probed with sera from patients with GC and DU. They identified 30 antigens and divided these into three groups: pathogen-unspecific, ulcer-associated, and gastric cancer-associated groups. Many immunoreactive proteins of H. pylori have been reported by 2D immunoblot studies, such as those mentioned above, but, as far as we are aware, only GroES has been confirmed to be associated with a particular disease (21). In this study, translation elongation factor EF-G (FusA) and flagellar hook protein were identified as H. pylori antigens for the first time. Furthermore we also identified three of the 11 proteins as candidate DU-related antigens by comparing the antigenic profile of H. pylori proteins from a DU strain probed with sera from patients with DU and GC. In addition to the preliminary data derived from 2D immunoblots, recombinant FusA, catalase (KatA), and urease ␣ subunit (UreA) were used to screen a larger sample size of clinical sera from patients with DU and GC, and the results confirmed the association with DU.

The physiological function of FusA is important for the translocation of peptidyl-tRNA from the A site to the P site within the ribosome (22). In addition, it is also an essential protein in the dissociation of the post-termination complex (23). In Staphylococcus aureus and most Gram-negative bacteria, such as Salmonella typhimurium, mutation of FusA causes fusidic acid resistance (24, 25). H. pylori UreA, the ␣ subunit of urease, is mainly found in the cytoplasm but is also presents at the cell surface and is secreted out of the bacterium (26, 27). H. pylori urease, an important H. pylori virulence factor, hydrolyzes urea into NH3 and CO2 and allows this bacterium to survive in the acidic gastric environment (28). Haas et al. (19) reported that UreA showed a significant difference in reactivity with sera from patients with gastroduodenal ulcer compared with gastritis. From our 2D immunoblot data, the frequency of reactivity with FusA and UreA in DU patients was 70 and 40 and 80% (n ⫽ 10), respectively. It was similar when a larger sample size was screened using the recombinant FusA (70.2%, n ⫽ 124), whereas it was lower using the recombinant UreA (44.4%, n ⫽ 124). H. pylori KatA was purified and characterized by Hazell et al. (29) and was found to be located in both the cytoplasm and periplasm (30). In addition, it is reported to be secreted out of H. pylori (31), and its catalase activity is essential for the survival of the bacterium at the phagocyte surface (32). Atanassov et al. (33) showed that the frequency of KatA seropositivity in patients with gastroduodenal ulcer was greater than 90% (n ⫽ 55), whereas that in patients with non-ulcer dyspepsia was 67% (n ⫽ 30) using an extract of a gastritis strain of H. pylori and 93% using the ATCC 43579 strain. In our study, as with the UreA data, the frequency of KatA seroreactivity in DU patients was 100% (n ⫽ 10) by 2D immunoblot


1023


FIG. 5. Development of a DU-related protein array for diagnosis. A, all proteins were spotted in quadruplicate at three different concentrations (100, 300, and 600 ␮g/ml) diluted with 0.14 –2 M urea in PBS. Each protein array contained human IgG at three dilutions (one normal serum was diluted with PBS to 1:1500, 1:1000, and 1:500) for interchip comparison. FlaA was used as a positive control antigen to monitor the experimental performance. The chips were incubated with sera from DU patients and normal individuals, and bound antibody was detected by the fluorescence of the Cy3-conjugated anti-human IgG (hIgG) antibodies. B, comparison of the serum reactions of six DU patients and seven normal individuals with the three DU-related antigens, FusA, KatA, and UreA. Student’s t test was used for statistical evaluation (*, p ⬍ 0.05). N, normal.

analysis but was 50.8% (n ⫽ 124) by larger sample size screening using recombinant KatA. Although KatA seropositivity was not associated more with DU than with GC, it improved the prediction of DU when combined with FusA and/or UreA (Table III). KatA is therefore also an important DU-related antigen. Our statistical analysis revealed that combinations of antigens improved the ability to distinguish between patients with DU and GC and normal individuals. The three-antigen combination gave excellent discrimination, and this phenomenon also implies the value of each antigen related to DU disease. However, heterogeneity of the host immune response to different H. pylori antigens may explain the relatively low occurrence of multiseropositivity in particular individuals in this study. For the preliminary examination of the risk of DU in the

1024


clinic, we suggest a rule considering positivity of two or three antigens as high risk to DU to compensate for this phenomenon. H. pylori infection can be diagnosed by noninvasive methods or by endoscopy. Serologic testing, a noninvasive method, is inexpensive and widely used for the diagnosis of H. pylori infection before treatment (34). Generally patients only agree to invasive endoscopy when they have alarming symptoms, such as stomachache, anemia, or gastrointestinal bleeding. Despite advances in therapies, gastrointestinal bleeding is still a cause of morbidity and mortality (35). Serologic tests are more acceptable and can be carried out in a routine physical examination. Following a finding of seropositivity for two or three antigens (i.e. FusA, KatA, and UreA), patients would be recommended to undergo endoscopic ex-


amination. We believe this concept will promote the early detection and treatment of DU disease. To apply the idea of diagnosis using multiple biomarkers, we constructed a protein array containing the three DU-related antigens and compared the antibody binding patterns of sera from DU patients and normal individuals. At each concentration of each of the three antigens, the signal intensity was higher with DU sera than with normal sera. Although the serum reaction using an antigen concentration of 100 ␮g/ml was able to distinguish the DU group from the controls, the use of a single cutoff value for a single protein concentration might lead to false positive or false negative results. We therefore used three concentrations of each antigen to monitor the trend of signal intensities. As long as two of the three signal intensities were higher than each cutoff value of protein concentrations, the sample was considered seropositive for a particular antigen. This results in fewer false positives. In conclusion, here we report a comparison of the reactivity of serum antibodies from DU and GC patients to the H. pylori proteome, leading to the identification of FusA, KatA, and UreA as DU-related antigens of H. pylori. We further demonstrate that the combination of these three antigens can be used as a biomarker as a preliminarily means of distinguishing patients with DU from those with GC. Moreover we have developed a DU-related protein array using the combination of these three antigens for the purpose of clinical diagnosis. Acknowledgments—We thank Sheng-Wei Lin for the assistance in nano-LC-MS/MS analysis and Pei-Yu Li for assistance in culturing H. pylori. * This work was supported in part by the Program for Excellence Research Teams from the Ministry of Education and Grant NSC 95-3112-B-002-016 from the National Science Council of the Republic of China. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www. mcponline.org) contains supplemental material. ¶¶ To whom correspondence should be addressed: Graduate Inst. of Biochemistry and Molecular Biology, College of Medicine, National Taiwan University, No. 1, Sec. 1, Jen-Ai Rd., Taipei 100, Taiwan. Tel.: 886-2-23123456 (ext. 8212); Fax: 886-2-23958814; E-mail: lupin@ha. mc.ntu.edu.tw. REFERENCES 1. Blaser, M. J. (1998) Helicobacter pylori and gastric diseases. BMJ 316, 1507–1510 2. Miwa, H., Go, M. F., and Sato, N. (2002) H. pylori and gastric cancer: the Asian enigma. Am. J. Gastroenterol. 97, 1106 –1112 3. Uemura, N., Okamoto, S., Yamamoto, S., Matsumura, N., Yamaguchi, S., Yamakido, M., Taniyama, K., Sasaki, N., and Schlemper, R. J. (2001) Helicobacter pylori infection and the development of gastric cancer. N. Engl. J. Med. 345, 784 –789 4. Blaser, M. J., and Atherton, J. C. (2004) Helicobacter pylori persistence: biology and disease. J. Clin. Investig. 113, 321–333 5. Dixon, M. F., Genta, R. M., Yardley, J. H., and Correa, P. (1996) Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. Am. J. Surg. Pathol. 20, 1161–1181

6. Portal-Celhay, C., and Perez-Perez, G. I. (2006) Immune responses to Helicobacter pylori colonization: mechanisms and clinical outcomes. Clin. Sci. (Lond.) 110, 305–314 7. Zevering, Y., Jacob, L., and Meyer, T. F. (1999) Naturally acquired human immune responses against Helicobacter pylori and implications for vaccine development. Gut 45, 465– 474 8. Ciociola, A. A., McSorley, D. J., Turner, K., Sykes, D., and Palmer, J. B. (1999) Helicobacter pylori infection rates in duodenal ulcer patients in the United States may be lower than previously estimated. Am. J. Gastroenterol. 94, 1834 –1840 9. Chu, K. M., Kwok, K. F., Law, S., and Wong, K. H. (2005) Patients with Helicobacter pylori positive and negative duodenal ulcers have distinct clinical characteristics. World J. Gastroenterol. 11, 3518 –3522 10. Chen, C. S., and Zhu, H. (2006) Protein microarrays. BioTechniques 40, 423– 429 11. Deinhofer, K., Sevcik, H., Balic, N., Harwanegg, C., Hiller, R., Rumpold, H., Mueller, M. W., and Spitzauer, S. (2004) Microarrayed allergens for IgE profiling. Methods 32, 249 –254 12. Lueking, A., Huber, O., Wirths, C., Schulte, K., Stieler, K. M., BlumePeytavi, U., Kowald, A., Hensel-Wiegel, K., Tauber, R., Lehrach, H., Meyer, H. E., and Cahill, D. J. (2005) Profiling of alopecia areata autoantigens based on protein microarray technology. Mol. Cell. Proteomics 4, 1382–1390 13. Wang, X., Yu, J., Sreekumar, A., Varambally, S., Shen, R., Giacherio, D., Mehra, R., Montie, J. E., Pienta, K. J., Sanda, M. G., Kantoff, P. W., Rubin, M. A., Wei, J. T., Ghosh, D., and Chinnaiyan, A. M. (2005) Autoantibody signatures in prostate cancer. N. Engl. J. Med. 353, 1224 –1235 14. Zhu, H., Hu, S., Jona, G., Zhu, X., Kreiswirth, N., Willey, B. M., Mazzulli, T., Liu, G., Song, Q., Chen, P., Cameron, M., Tyler, A., Wang, J., Wen, J., Chen, W., Compton, S., and Snyder, M. (2006) Severe acute respiratory syndrome diagnostics using a coronavirus protein microarray. Proc. Natl. Acad. Sci. U. S. A. 103, 4011– 4016 15. Utt, M., Nilsson, I., Ljungh, A., and Wadstrom, T. (2002) Identification of novel immunogenic proteins of Helicobacter pylori by proteome technology. J. Immunol. Methods 259, 1–10 16. Kao, S. H., Su, S. N., Huang, S. W., Tsai, J. J., and Chow, L. P. (2005) Sub-proteome analysis of novel IgE-binding proteins from Bermuda grass pollen. Proteomics 5, 3805–3813 17. Peterson, W. L. (1991) Helicobacter pylori and peptic ulcer disease. N. Engl. J. Med. 324, 1043–1048 18. Kimmel, B., Bosserhoff, A., Frank, R., Gross, R., Goebel, W., and Beier, D. (2000) Identification of immunodominant antigens from Helicobacter pylori and evaluation of their reactivities with sera from patients with different gastroduodenal pathologies. Infect. Immun. 68, 915–920 19. Haas, G., Karaali, G., Ebermayer, K., Metzger, W. G., Lamer, S., ZimnyArndt, U., Diescher, S., Goebel, U. B., Vogt, K., Roznowski, A. B., Wiedenmann, B. J., Meyer, T. F., Aebischer, T., and Jungblut, P. R. (2002) Immunoproteomics of Helicobacter pylori infection and relation to gastric disease. Proteomics 2, 313–324 20. Mini, R., Bernardini, G., Salzano, A. M., Renzone, G., Scaloni, A., Figura, N., and Santucci, A. (2006) Comparative proteomics and immunoproteomics of Helicobacter pylori related to different gastric pathologies. J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 833, 63–79 21. Lin, Y. F., Wu, M. S., Chang, C. C., Lin, S. W., Lin, J. T., Sun, Y. J., Chen, D. S., and Chow, L. P. (2006) Comparative immunoproteomics of identification and characterization of virulence factors from Helicobacter pylori related to gastric cancer. Mol. Cell. Proteomics 5, 1484 –1496 22. Selmer, M., Al-Karadaghi, S., Hirokawa, G., Kaji, A., and Liljas, A. (1999) Crystal structure of Thermotoga maritima ribosome recycling factor: a tRNA mimic. Science 286, 2349 –2352 23. Rao, A. R., and Varshney, U. (2001) Specific interaction between the ribosome recycling factor and the elongation factor G from Mycobacterium tuberculosis mediates peptidyl-tRNA release and ribosome recycling in Escherichia coli. EMBO J. 20, 2977–2986 24. Besier, S., Ludwig, A., Brade, V., and Wichelhaus, T. A. (2003) Molecular analysis of fusidic acid resistance in Staphylococcus aureus. Mol. Microbiol. 47, 463– 469 25. Johanson, U., and Hughes, D. (1994) Fusidic acid-resistant mutants define three regions in elongation factor G of Salmonella typhimurium. Gene (Amst.) 143, 55–59


1025


26. Hawtin, P. R., Stacey, A. R., and Newell, D. G. (1990) Investigation of the structure and localization of the urease of Helicobacter pylori using monoclonal antibodies. J. Gen. Microbiol. 136, 1995–2000 27. Vanet, A., and Labigne, A. (1998) Evidence for specific secretion rather than autolysis in the release of some Helicobacter pylori proteins. Infect. Immun. 66, 1023–1027 28. Dunn, B. E., Cohen, H., and Blaser, M. J. (1997) Helicobacter pylori. Clin. Microbiol. Rev. 10, 720 –741 29. Hazell, S. L., Evans, D. J., Jr., and Graham, D. Y. (1991) Helicobacter pylori catalase. J. Gen. Microbiol. 137, 57– 61 30. Harris, A. G., and Hazell, S. L. (2003) Localisation of Helicobacter pylori catalase in both the periplasm and cytoplasm, and its dependence on the twin-arginine target protein, KapA, for activity. FEMS Microbiol. Lett. 229, 283–289

1026


31. Mori, M., Suzuki, H., Suzuki, M., Kai, A., Miura, S., and Ishii, H. (1997) Catalase and superoxide dismutase secreted from Helicobacter pylori. Helicobacter 2, 100 –105 32. Ramarao, N., Gray-Owen, S. D., and Meyer, T. F. (2000) Helicobacter pylori induces but survives the extracellular release of oxygen radicals from professional phagocytes using its catalase activity. Mol. Microbiol. 38, 103–113 33. Atanassov, C., Pezennec, L., d’Alayer, J., Grollier, G., Picard, B., and Fauchere, J. L. (2002) Novel antigens of Helicobacter pylori correspond to ulcer-related antibody pattern of sera from infected patients. J. Clin. Microbiol. 40, 547–552 34. Suerbaum, S., and Michetti, P. (2002) Helicobacter pylori infection. N. Engl. J. Med. 347, 1175–1186 35. Kashiwagi, H. (2005) Ulcers and gastritis. Endoscopy 37, 110 –115